Dylan Mackenzie2, Elisa Fevola1, Mark Penney1
1AstraZeneca, 2Coventry University
Introduction/Objectives: T-cell engagers (TCEs) are promising targeted immunotherapeutic agents, which can redirect the immune system’s T cells to recognize and kill cancer cells. Despite their potential, the clinical development of TCEs presents many challenges, mainly due to their complex Mechanism of Action, which relies on the formation of an active trimer between drug, tumour cells and T cells to drive tumour cell killing, and to their narrow therapeutic window, with potential risk of cytokine release syndrome (CRS) even at low doses. One challenge is selecting a First-in-Human (FIH) dose that is both safe and minimally efficacious, to potentially cut the number of cohorts exposed to sub-efficacious doses in Phase 1 clinical trials. In this context, mechanistic modelling is emerging as a powerful tool for selecting the FIH starting dose. In particular, a trimer-based minimal anticipated biological effect level (MABEL) approach could potentially lead to higher starting doses than the traditional MABEL approach, while still being safe. In this work, we develop a mechanistic framework for the estimation of a trimer-based MABEL dose for blinatumomab, a TCE directed against CD19 for B cell malignancies. The obtained FIH starting dose is contextualized by considering the efficacy and safety observations emerged for blinatumomab in the clinic. Moreover, we evaluate the impact of using different tumour tissues and different trimer normalization approaches on the estimated FIH starting dose. Methods: A modelling framework composed of an in vitro and an in vivo trimer model was developed, based on the models reported in [1,2,3]. The in vitro model was fitted to in vitro cytotoxicity data obtained from the literature [1], and it was used to identify the in vitro trimer EC50, defined as the trimer concentration leading to 50% of the maximal activity. The in vivo model was then used to simulate the trimer formation in human, in the relevant tumour tissues. A FIH dose was then estimated as the dose leading to a tumour trimer concentration equal to the in vitro trimer EC50. To estimate the impact of using different tumour tissues to drive the selection of the FIH starting dose, in vivo trimer concentrations were estimated both in the blood and in the bone marrow [1] to match the in vitro trimer EC50. Moreover, the translatability of a trimer-based metric from in vitro to in vivo could be impacted by the large difference between the two scenarios in terms of E:T ratios and cell densities. For this reason, we assessed the impact on the estimated FIH dose of expressing the trimer EC50 either as molar concentration, or in terms of number of trimers formed per tumour cell, or per T cell. Results: The mechanistic modelling framework was tested on blinatumomab data [1]. The estimated FIH doses ranged from 0.1 to 30 µg/m2/day depending on the tissue considered (blood or bone marrow), and the trimer normalization used (molar concentration, Trimer/Tumor cell, or Trimer/T cell). All these combinations (apart from trimer molar concentration in the bone marrow) led to higher starting doses than the FIH dose of 0.5 µg/m2/day used for blinatumomab. Additionally, all doses were below the maximum tolerated dose of 60 µg/m2/day, with some above 5 µg/m2/day, at which clinical data showed complete depletion of peripheral B-lymphocytes [4]. The use of blood concentration consistently led to higher starting doses than bone marrow, likely due to the lower E:T ratio and cell densities in blood compared to bone. The use of different trimer normalizations resulted in different predicted starting doses, which were the highest when using trimers/T cells and the lowest when using trimer molar concentration. Conclusions: A trimer-based MABEL approach was retrospectively evaluated for the FIH starting dose estimation of blinatumomab. The FIH doses obtained with the proposed approach were in most cases higher than the FIH dose used for blinatumomab, and always in the range of doses that were tolerated in the clinic, confirming the potential benefits of adopting a mechanistic trimer-based approach for selecting the starting dose of T-cell engagers. The use of different tumour tissues and trimer normalizations resulted in a large variability of the estimated doses, highlighting the impact that these small changes in the modelling framework can have on the selected starting dose.
[1] Jiang, Xiling, et al. “Development of a target cell-biologics-effector cell (TBE) complex-based cell killing model to characterize target cell depletion by T cell redirecting bispecific agents.” MAbs. Vol. 10. No. 6. Taylor & Francis, 2018. [2] Chen, X., et al. “Mechanistic projection of first-in-human dose for bispecific immunomodulatory P-Cadherin LP-DART: an integrated PK/PD modeling approach.” Clinical Pharmacology & Therapeutics 100.3 (2016): 232-241. [3] Betts, Alison, et al. “A translational quantitative systems pharmacology model for CD3 bispecific molecules: application to quantify T cell-mediated tumor cell killing by P-cadherin LP DART®.” The AAPS journal 21 (2019): 1-16. [4] Hijazi, Youssef et al. “Pharmacokinetic and Pharmacodynamic Relationship of Blinatumomab in Patients with Non-Hodgkin Lymphoma.” Current clinical pharmacology vol. 13,1 (2018): 55-64. doi:10.2174/1574884713666180518102514
Reference: PAGE 33 (2025) Abstr 11432 [www.page-meeting.org/?abstract=11432]
Poster: Methodology - New Modelling Approaches